Provided are a fisheye camera calibration system, method and an electronic device. The system includes a hemispherical target, a fisheye camera and an electronic device. The hemispherical target includes a hemispherical inner surface and multiple markers provided on the hemispherical inner surface. The fisheye camera is used for photographing the hemispherical target and acquiring a target image, where the hemispherical target and the multiple markers provided on the hemispherical inner surface are captured in the target image. The electronic device is used for acquiring initial values of k.sub.1, k.sub.2, k.sub.3, k.sub.4, k.sub.5, u.sub.0, v.sub.0, m.sub.u and m.sub.v, and using a Levenberg-Marquardt algorithm to optimize the initial values of k.sub.1, k.sub.2, k.sub.3, k.sub.4, k.sub.5, u.sub.0, v.sub.0, m.sub.u and m.sub.v, so as to determine imaging model parameters of the fisheye camera.

Systems and techniques are described for large field of view digital imaging. A device's first image sensor captures a first image based on first light redirected from a first path onto a redirected first path by a first light redirection element, and the device's second image sensor captures a second image based on second light redirected from a second path onto a redirected second path by a second light redirection element. A virtual extension of the first path beyond the first light redirection element can intersect with a virtual extension of the second path intersect beyond the second light redirection element. The device can modify the first image and second image using perspective distortion correction, and can generate a combined image by combining the first image and the second image. The combined image can have a larger field of view than the first image and/or the second image.

Even if a generated wide-field image includes residual local misalignment, this charged particle microscope device can prompt for user input to correct such local misalignment, and can regenerate, on the basis of the user input, a wide-field image that includes little misalignment even in local areas of the overlap regions thereof. A charged particle microscope according to the present invention captures a plurality of images in such a way that each captured image has overlap regions that are to be overlapped with the overlap regions of captured images adjacent to that captured image, wherein an image processing unit: sets a pair of corresponding points in respective overlap regions of each two adjacent captured images; sets predetermined constraint conditions for each captured image; calculates the amounts of misalignment between the plurality of captured images on the basis of the set pairs of corresponding points and the set constraint conditions; connects the plurality of captured images to one another after correcting the misalignment between these captured images on the basis of the calculated amounts of misalignment, thereby generating a single wide-field image; calculates, for each of a plurality of local areas set in the overlap regions of each two adjacent captured images, a degree of reliability for the connection between these captured images; and notifies a user of either each found low reliability local area or the overlap region including that low reliability local area, as well as the set pairs of corresponding points and the set constraint conditions.

A processor-implemented method with data processing using a neural network includes: determining a first translated image by translating a first image based on a second image, the first image and a second image that having different distortions, such that a distortion of the first image corresponds to a distortion of the second image; determining a first retranslated image by translating the first translated image such that a distortion of the first translated image corresponds to a distortion of the first image; and training a first deformation field generator configured to determine a first relative deformation field that represents a relative deformation from the first image to the second image and a second deformation field generator configured to determine a second relative deformation field that represents a relative deformation from the second image to the first image, based on a loss between the first retranslated image and the first image.

In one embodiment, a method includes a computer system accessing a curvilinear image captured using a camera lens, generating multiple rectilinear images from the curvilinear image based at least in part on one or more calibration parameters associated with the camera lens, identifying semantic information in one or more of the rectilinear images by processing each of the multiple rectilinear images using a machine-learning model configured to identify semantic information in rectilinear images, and identifying semantic information in the curvilinear image based on the identified semantic information in the one or more rectilinear images.

An endoscope apparatus includes a processor configured to carry out a reversal process on a side image of a subject picked up with an image pickup device, the reversal process reversing an order of a plurality of pixels located in the side image and arranged from a predetermined position in a radial direction, and cause a display apparatus to display a front image and the side image of the subject picked up with the image pickup device as a display image. The processor adjusts the timing at which the front image is displayed in such a way that the front image acquired simultaneously with the side image is displayed at a timing that accords with the timing at which the side image having undergone the reversal process is displayed on the display apparatus.

Systems and methods are provided to generate an object detection representation of a candidate object based on sensor data representing a captured image of an environment surrounding the vehicle. A determination is made as to whether, whether the candidate object is an outlier based on the object detection representation. In response to determining the candidate object is not an outlier, the candidate object is validated as an object, a proximity distance of the object to the vehicle is determined, and the proximity distance of the object is sent as output.

System, methods, and other embodiments detecting and localizing objects within an image in a wide-view format using a synthetic representation. The method includes converting a real image in a wide-view format to a synthetic representation using a style model, wherein the synthetic representation depicts a distorted view of an object. The method also includes identifying features of the object using an extraction model that distinguishes different scales of the synthetic representation and a simulated scene to define structures associated with the distorted view. The method also includes detecting the object using a decoder model that identifies an attribute and a bounding box of the object from the features. The method also includes executing a task using the attribute and the bounding box to localize the object in the simulated scene.

An image recognizing device includes an image storing portion, a cylindrical distortion correcting portion, a vertical edge extracting portion; a column candidate extracting portion, a pole candidate evaluating portion, a pole foot position setting portion, a movement distance acquiring portion, a detected distance difference calculating portion, and a pole identifying portion. When the vehicle is moving toward a pole candidate, the movement distance acquiring portion acquires a movement distance moved by the vehicle during a prescribed time interval, and the detected distance difference calculating portion calculates a detected distance difference between a starting detected distance and an ending detected distance for the prescribed time interval. The pole identifying portion identifies that a pole candidate wherein the absolute value of the difference between the movement distance and the detected distance difference is less than a threshold value as being a pole that has a pole foot position that contacts the ground.

Hyper-hemispherical images may be combined to generate a rectangular projection of a spherical image having an equatorial stitch line along of a line of lowest distortion in the two images. First and second circular images are received representing respective hyper-hemispherical fields of view. A video processing device may project each circular image to a respective rectangular image by mapping an outer edge of the circular image to a first edge of the rectangular image and mapping a center point of the circular image to a second edge of the first rectangular image. The rectangular images may be stitched together along the edges corresponding to the outer edge of the original circular image.